Báo cáo Y học: Expression of recombinant murine pregnancy-associated plasma protein-A (PAPP-A) and a novel variant (PAPP-Ai) with differential proteolytic activity pot

The human intron flanked by these exons does not encode a homologous corresponding insert, which is unique to the mouse.The overall sequence identity between murine and human PAPP-A is 91

Expression of recombinant murine pregnancy-associated plasma

protein-A (PAPP-A) and a novel variant (PAPP-Ai) with differential proteolytic activity

Rikke Søe1, Michael T Overgaard1, Anni R Thomsen1, Lisbeth S Laursen1, Inger M Olsen1,

Lars Sottrup-Jensen1, Jesper Haaning1, Linda C Giudice2, Cheryl A Conover3and Claus Oxvig1

1

Department of Molecular and Structural Biology, Science Park, University of Aarhus, Denmark;2Department of Gynecology and Obstetrics, Stanford University, Stanford, CA, USA;3Endocrine Research Unit, Mayo Clinic and Foundation, Rochester, USA

Murine pregnancy-associated plasma protein-A (PAPP-A)

cDNA encoding a 1545 amino-acid protein has been cloned

We have also identified and cloned cDNA that encodes a

novel variant of PAPP-A, PAPP-Ai, carrying a 29-residue

highly basic insert.The point of insertion corresponds to a

junction between two exons in the human PAPP-A gene

The human intron flanked by these exons does not encode a

homologous corresponding insert, which is unique to the

mouse.The overall sequence identity between murine and

human PAPP-A is 91%, and murine PAPP-A contains

sequence motifs previously described in the sequence of

human PAPP-A.Through expression in mammalian cells,

we show that murine PAPP-A and PAPP-Ai are active

metalloproteinases, both capable of cleaving insulin-like

growth factor binding protein (IGFBP)-4 and -5.Cleavage

of IGFBP-4 is dramatically enhanced by the addition of

IGF, whereas cleavage of IGFBP-5 is slightly inhibited by

IGF, as previously established with human

PAPP-A.Sur-prisingly, however, quantitative analyses demonstrate that

the murine PAPP-Ai cleaves IGFBP-4 very slowly

com-pared to PAPP-A, even though its ability to cleave IGFBP-5

is unaffected by the presence of the insert.By RT-PCR

analysis, we find that both variants are expressed in several tissues.The level of mRNA in the murine placenta does not exceed the levels of other tissues analyzed.Furthermore, the IGFBP-4-proteolytic activity of murine pregnancy serum is not elevated.This is in striking contrast to the increase seen

in human pregnancy serum, and the expression of PAPP-A

in the human placenta, which exceeds other tissues at least 250-fold.Interestingly, the position of the insert of PAPP-Ai, within the proteolytic domain, lies in close proximity to the cysteine residue, which in human PAPP-A forms a disulfide bond with the proform of eosinophil major basic protein (proMBP).ProMBP functions as a proteinase inhibitor in the PAPP-A–proMBP complex, but whether any mechan-istic parallel on regulation of proteolytic activity can be drawn between the insert of PAPP-Ai and the linkage to proMBP is not known.Importantly, these data support the development of the mouse as a model organism for the study

of PAPP-A, which must take into account the differences between the mouse and the human

Keywords: metalloproteinase; metzincin; insulin-like growth factors; IGF binding proteins; pregnancy proteins

Insulin-like growth factors (IGF)-I and -II are established

regulators of growth in many systems [1].Their activity is

modulated by IGF binding proteins (IGFBPs), six of which

are known [2,3].The IGFBPs bind IGF-I and -II with high

affinities, but proteolytic cleavage in the central region of an

IGFBP causes loss of its affinity for IGF.Thus, proteolysis

can be a prerequisite for the exertion of IGF activities [4]

Human pregnancy-associated plasma protein-A (PAPP-A)

was recently identified as a proteinase specific for IGFBP-4

[5] and IGFBP-5 [6].Interestingly, its cleavage of IGFBP-4

is dramatically enhanced by the presence of IGF, whereas the cleavage of IGFBP-5 is slightly reduced [6]

PAPP-A is a glycoprotein of 1547 residues [7], originally isolated from the serum of pregnant women, but recently also described in a number of human systems and shown to

be secreted from fibroblasts [5], osteoblasts [5,8], vascular smooth muscle cells [9,10], and ovarian granulosa cells [11,12].In pregnancy, PAPP-A is synthesized in the human placenta [13].It reaches high levels in third trimester serum ( 50 mgÆL)1) [14], where it circulates as a disulfide bound

2 : 2 complex of 500 kDa with the proform of eosinophil major basic protein (proMBP) [15,16].The mature form of the 206-residue proMBP, the 117-residue MBP, has a calculated isoelectric point of 11 and is thus extremely basic MBP is cytotoxic and is found in granules of the eosinophil leukocyte, from which it is secreted as a defense mechanism

of the immune system [17].It has recently been demonstra-ted that in the complex with PAPP-A, proMBP functions as

a proteinase inhibitor of unknown mechanism [18].How-ever, human pregnancy serum does show proteolytic activity against IGFBP-4, because the amount of circulating

Correspondence to C.Oxvig, Department of Molecular and Structural

Biology, Science Park, University of Aarhus, Gustav Wieds Vej 10C,

DK-8000 Aarhus C, Denmark.

Fax: + 45 86123178, E-mail: co@mbio.aau.dk

Abbreviations: PAPP-A, pregnancy-associated plasma protein-A;

PAPP-Ai, variant of PAPP-A with 29-residue insert in the proteolytic

domain; IGF, insulin-like growth factor; IGFBP, insulin-like growth

factor binding protein.

Enzyme: pregnancy-associated plasma protein-A (EC 3.4.24.-).

(Received 23 November 2001, revised 11 March 2002,

accepted 15 March 2002)

PAPP-A is high (compared to PAPP-A levels elsewhere),

and because a minor fraction (< 1%) is found as a

noncomplexed PAPP-A dimer of about 400 kDa [18]

PAPP-A belongs to the metzincin superfamily of

metal-loproteinases [19,20], a diverse group of zinc peptidases

comprised of five families: the astacins (e.g bone

morpho-genetic protein-1), the reprolysins or adamalysins (snake

venom proteinases, ADAMs), the serralysins (bacterial

proteinases), the matrix metalloproteinases (MMPs or

matrixins), and the pappalysins [21].In addition to

PAPP-A, the latter includes PAPP-A2, a recently discovered

human homologue of PAPP-A showing 45% sequence

identity to PAPP-A [22].All metzincins contain an

elonga-ted zinc-binding motif (HEXXHXXGXXH), which

coordinates the catalytic zinc ion of the active site.In

addition, they have a strictly conserved Met-residue, located

in the sequences at a variable distance (7–63 residues) to the

zinc-binding site [21], but in an invariable so-called Met-turn

in the three-dimensional structures [20].PAPP-A and

PAPP-A2 further contain three lin-notch motifs (LNR1-3)

and five short consensus repeats (SCR1-5) [21]

Proteolysis of IGFBP-4 or -5 has been reported in

conditioned media from cultures of rat B104 neuroblastoma

cells [23], murine osteoblasts [24–26], rat ovarian granulosa

cells [27], and rat vascular smooth muscle cells [28]

However, whether PAPP-A exists in mouse as an active

enzyme is unknown

We have cloned the cDNAs encoding murine PAPP-A

and a novel variant, PAPP-Ai, not known in humans, and

we have shown that mRNAs encoding both species are

expressed in several murine tissues.Recombinant expression

in mammalian cells allowed biochemical characterization of

murine PAPP-A and PAPP-Ai.Development of the mouse

as a model organism for the study of PAPP-A is an

important goal for understanding PAPP-A as a unique

metalloproteinase

E X P E R I M E N T A L P R O C E D U R E S

Cloning of cDNAs encoding murine PAPP-A and PAPP-Ai

Overlapping, murine cDNA clones encoding murine

PAPP-A of 1545 residues and a 1574-residue variant, PPAPP-APP-PAPP-Ai,

were isolated using standard procedures.Nucleotides 1–3 of

the deposited sequences (AF439513, PAPP-A and

AF439514, PAPP-Ai) encode residue 1 (Fig.1).Nucleotide

numbers below refer to the cDNA sequence of AF439513,

and residue numbers refer to its translated amino acid

sequence (Fig.1), unless otherwise specified In brief, a

murine cDNA library constructed in the lambda

ZAP-CMV vector (Stratagene) from murine placentas of a

17-day-pregnant mouse was screened using a32P-labeled

human PAPP-A cDNA probe derived from pPA1 [7].From

a total of > 500 000 plaques, only two clones were found to

contain murine PAPP-A cDNA.Following rescue from the

lambda vector, the two clones were contained in the

pBK-CMV phagemid vector.One clone (E11) contained

nucleo-tides 1493–4638 of the murine PAPP-A cDNA sequence,

coding for residues 499–1545 of murine PAPP-A, followed

by a stop codon, and a 3¢-UTR sequence of  400

nucleotides.The sequence of the other clone was contained

within E11.No further sequence was obtained in

subse-quent rounds of re-hybridization using probes derived from E11

Using two rounds of RT-PCR with SuperTaq DNA polymerase (HT Biotechnologies), the remaining nucleo-tides of the murine PAPP-A cDNA sequence were obtained

In the first round, cDNA was synthesized from term placental RNA using a primer derived from E11 (nucleo-tides 1695–1708).PCR was carried out with the 3¢ primer derived from E11 (nucleotides 1664–1682), and the 5¢ primer derived from the human PAPP-A cDNA sequence (nucle-otides 3939–3959 of NM_002581).The resulting PCR product was cloned into pCR 2.1-TOPO (Invitrogen) The clone F2 contained nucleotides 478–1682, encoding residues 161–560.A variant, clone F2i, contained the same sequence,

in addition to an in-frame insert of 87 nucleotides (between nucleotides 1232 and 1233 of AF439513, corresponding to

an insertion between amino acid residues 411 and 412).The PAPP-A variant carrying this insert is denoted PAPP-Ai Similarly, clone F1 was obtained using an RT-primer derived from F2 (nucleotides 847–866), a 3¢ primer derived from F2 (nucleotides 507–526), and a 5¢ primer derived from the human cDNA sequence (nucleotides 3438–3453 of NM_002581).F1 contained nucleotides 1–525, encoding residues 1–175.Several independent clones (of F2, F1, and F2i) resulting from this PCR-based procedure were isolated and found to be identical

Generation of expression constructs Expression constructs encoding full-length murine PAPP-A and PAPP-Ai were made using a signal peptide previously used for the expression of human PAPP-A [18].First, T1218

of F2 was substituted with a G by overlap extension PCR [29], using outer primers derived from the vector (nucleo-tides 212–233 and 404–426 of pCR 2.1 TOPO), and overlapping, inner primers derived from the PAPP-A cDNA (nucleotides 1209–1232 and 1205–1226).This cre-ated a silent BspEI site at residue Pro406 that will facilitate future mutagenesis.The resulting PCR product was diges-ted with PstI (nucleotide 1502) and XbaI (derived from the vector), ligated into the PstI–ClaI fragment (nucleotides 1502–1845) of E11, and cloned into the XbaI/ClaI sites of pBluescript II SK+ (Stratagene) to obtain pB-F2PC (con-taining nucleotides 478–1845 encoding residues 161–615 of mPAPP-A)

Using the human PAPP-A expression construct (pcDNA3.1-PAPP-A [18]) as a template and primers derived from pcDNA3.1+ (Invitrogen) (nucleotides 792– 812) and pcDNA3.1-PAPP-A/F1 (5¢-GCCCCTCGCCGC GCTCGAGGCCG-3¢), a PCR product containing a HindIII site followed by nucleotides encoding the signal peptide (MKDSCITVMAMALLSGFFFFAPASS) and murine PAPP-A residues 1–4 was obtained.Using F1 as a template, a primer derived in part from nucleotides 1–13 (5¢-GGCCTCGAGCGCGGCGAGGGGCG-3¢), and a primer derived from the vector (nucleotides 212–233 of pCR 2.1 TOPO), an overlapping PCR product was generated.The two PCR products were then used in an overlap reaction, and the resulting product cloned into pCR-Blunt II-TOPO (Invitrogen) to obtain pCR-sp.-F1 (encoding the signal peptide and residues 1–175 of mPAPP-A)

Next, using pB-F2PC and pCR-sp.-F1 as templates,

outer primers derived from the vector sequences

(nucleo-tides 625–645 of pBluescript II SK+ and nucleo(nucleo-tides 792–

812 of pcDNA3.1+ (Invitrogen), and inner primers derived

from the murine PAPP-A cDNA sequence (nucleotides

478–499 and 504–524), a PCR product encoding the signal

peptide and residues 1–615 was generated and cloned into pCR-Blunt II-TOPO.Finally, the HindIII–ClaI fragment was excised from this construct, ligated to the ClaI–BamHI fragment (encoding residues 616–1545) of E11, and cloned into pcDNA3.1+, to finally obtain pcDNA3.1-mPA Using F2i rather than F2, pBF2iPC, and further

pcDNA3.1-Fig 1 Alignment of the murine (mPA, 1545 residues) and human (hPA, 1547 residues) A sequences A variant of the murine protein,

PAPP-Ai, carries a 29-residue, basic insert whose amino acid sequence and position within the proteolytic domain (between residues 411 and 412) is emphasized.The extent in primary structure of the proteolytic domain, as recently defined [21], is indicated by the shading of residues 270–581.The human sequence [7] (GenBank accession number X68280) is shown only where different from the murine sequence.Murine PAPP-A contains 91.1% residues which are also found in the human protein; all of the 82 cysteine residues are conserved The elongated zinc-binding site (residues 480–490) and the Met-turn residues (552–556) [21] are shown in bold and underlined.Other defined stretches of amino acids are three lin-notch motifs (LNR1-3) and five short consensus repeats (SCR1-5).The PAPP-A cysteine residue known to be engaged in disulfide bonding to proMBP in the human PAPP-A–proMBP complex, the human Cys381 [15], is pointed out.

mPAi, was generated in parallel by the same procedure.All

PCRs were carried out using Pfu DNA polymerase

(Stratagene), and the final constructs were verified by

sequence analysis

Tissue culture and expression of recombinant proteins

Human embryonic kidney 293T cells (293tsA1609neo) [30]

were maintained in high-glucose Dulbecco’s modifies

Eagle’s medium supplemented with 10% fetal bovine

serum, 2 mM glutamine, nonessential amino acids, and

gentamicin (Life Technologies).Cells were plated onto 6-cm

tissue culture dishes, and were transfected 18 h later by

calcium phosphate coprecipitation [31] using 10 lg of

plasmid DNA prepared by QIAprep Spin Kit (Qiagen)

After a further 48 h, the supernatants were harvested and

cleared by centrifugation.Expression plasmids containing

murine PAPP-A, murine PAPP-Ai, or human PAPP-A [18]

were used for transfection.The level of human PAPP-A in

culture supernatants, determined by ELISA specific for

human PAPP-A, was about 5 lgÆmL)1 or 25 nM (of

200 kDa PAPP-A monomer), as reported previously [18]

No immunoassay is available for murine PAPP-A, but

based on quantitative comparison (cf.below) of the activity

against both IGFBP-4 and IGFBP-5 (not shown), it can be

suggested that supernatant from cells transfected murine

PAPP-A cDNA also contain 25 nM.However,

superna-tant from cells transfected with murine PAPP-Ai cDNA

showed very little activity against IGFBP-4 and less activity

against IGFBP-5, compared to murine PAPP-A.The levels

of murine PAPP-A and PAPP-Ai were therefore compared

by Western blotting using polyclonal rabbit antibodies

raised against human PAPP-A/proMBP complex [16], and

enhanced chemiluminescence (ECL Plus, Amersham)

Although raised against human protein, this preparation

of antibodies did result in a signal in Western blotting

experiments.A threefold difference in expression levels was

found (not shown) that was subsequently adjusted for

Expression of recombinant binding proteins was similarly

carried out, and purification was performed as recently

described for IGFBP-4 [21] and IGFBP-5 [22]

Measurement of proteolytic activity against IGFBP-4

and -5

All digests were carried out in 100 mMNaCl, 1 mMCaCl2,

50 mMTris, pH 7.5 using purified and iodinated IGFBP-4

[21] and IGFBP-5 [22].The reaction mixtures were analyzed

by nonreducing SDS/PAGE (16%) followed by

autoradi-ography.The material loaded per lane (10 lL) contained

25 000 c.p.m ( 2.5 ng or 7 nM) of radiolabeled binding

protein.All reactions were incubated at 37C (up to 72 h)

as specified in the text.Both of the binding proteins were

expressed as C-terminally tagged proteins causing the

PAPP-A cleavage products to comigrate, as detailed

previously [6]

For qualitative assays (Fig.2), unlabeled, purified

IGFBP-4 or -5 was added to a final concentration of

30 nM.Standard serum containing media from cells

trans-fected with empty vector or PAPP-A/PAPP-Ai cDNA were

used ( 0.1 nM proteinase), and 40 nM IGF-II (Bachem)

was added to some reactions as specified.Quantitative

assays (Fig.4A,B), recently developed [6], were carried out

after addition of 144 /130 nM unlabeled IGFBP-4/)5,

167 nMIGF-II, and equal amounts of proteinase (0.1 nM) contained in culture supernatants.For each experiment, one reaction of 120 lL was set up, from which samples of 10 lL were taken out at selected time points, stopped by the addition of 5 mM EDTA, and frozen.Following SDS/ PAGE, the degree of cleavage was determined by measuring band intensities with a PHOSPHORIMAGER (Molecular Dynamics) [6,21].The background signal from a control reaction using medium from mock-transfected cells was subtracted, and the degree of cleavage was plotted as a function of time.For evaluation of IGFBP-4 proteolytic activity in sera, blood was drawn from nonpregnant and pregnant (18 days) mice, and from nonpregnant and pregnant women (at term).Serum (0.5 lL) was used in each reaction along with 40 nMof added IGF-II

Analysis of tissue expression by RT-PCR Selected tissues from nonpregnant and pregnant mice were frozen in liquid nitrogen.Individual tissues (approximately

30 mg) was homogenized, further processed using QIA-shredder (Qiagen), and RNA was prepared using RNeasy

Fig 2 Proteolyticactivity of recombinant murine A and

PAPP-Ai against IGFBP-4 and -5 (A) Radiolabeled IGFBP-4 was incubated (24 h) with medium from mock-transfected cells (lane 1), with medium from cells transfected with murine PAPP-A cDNA (lanes 2–3), with murine PAPP-Ai cDNA (lanes 4–5), or with human PAPP-A cDNA (lane 6).Below each lane the absence (–) or presence (+) of 40 n M

added IGF-II is indicated.B: Similar experiment carried out with radiolabeled IGFBP-5 (except IGF-II was not added in lane 6, as indicated).The positions of molecular mass markers, and the positions

of intact and cleaved IGFBP-4 and -5 are indicated.The C-terminal tag on both of the binding protein causes their PAPP-A cleavage products to comigrate, and thus appear as one band, as detailed pre-viously [6].

Mini Kit (Qiagen).First strand cDNA was synthesized with

Thermoscript RT-PCR System (Life Technologies) and a

primer derived from the murine PAPP-A sequence

(nucleo-tides 1425–1448 of AF439513).PCR primers were chosen

that would amplify a nucleotide stretch spanning the cDNA

encoding the 29-residue insert of PAPP-Ai (nucleotides

1117–1137 and nucleotides 1275–1293 of AF439513),

resulting in a 177-bp product specific for cDNA derived

from PAPP-A mRNA, and 264-bp product specific for

PAPP-Ai mRNA.To analyze for the presence of a

corresponding insertion of human PAPP-A, the same

analysis was performed on human tissue using an RT

primer (nucleotides 4885–4908 of NM_002581) and two

equivalent PCR primers (nucleotides 4577–4597 and

nucleo-tides 4735–4753 of NM_002581) derived from the human

PAPP-A sequence.Reactions (33 cycles of PCR) were

carried out using SuperTaq DNA polymerase (HT

Bio-technologies)

R E S U L T S

Isolation of cDNAs encoding murine PAPP-A

and PAPP-Ai

A cDNA library prepared from murine placenta was

screened with a nucleotide probe encoding human PAPP-A

Only two clones, covering the C-terminal two thirds of the

sequence of murine PAPP-A, were found by hybridization

The remaining sequence was obtained by an RT-PCR based

procedure, as detailed above

The deduced amino-acid sequence of murine PAPP-A

contains 1545 residues, 137 of which differ from human

PAPP-A.Thus, PAPP-A is highly conserved with 91.1%

identical residues between the two species (Fig.1).Extended

stretches of identical residues occur, but positions that

deviate appear evenly distributed.Importantly, all of the 82

cysteine residues are conserved

Of particular interest, we have demonstrated the existence

of an mRNA species encoding a variant of PAPP-A with 29

TR) inserted in the proteolytic domain, between Arg411

and Ala412 (Fig 1).We denoted this variant PAPP-Ai.The

point of insertion does not interrupt any of the predicted

secondary-structure elements of PAPP-A, but is located

next to the N-terminal end of the pappalysin-specific a helix

H-ii, between the canonical b strands S3 and S4 of the

metzincins [21].With 11 basic (six Lys and five Arg) and two

acidic (both Glu) residues, the inserted stretch is highly

basic.The nucleotide sequence encoding murine PAPP-A

has been deposited in the GenBank database under the

accession number AF439513, and the sequence encoding

PAPP-Ai under the accession number AF439514

Does the human PAPP-A gene encode a similar insert?

The relevant portion of the human PAPP-A amino-acid

sequence is encoded by the genomic sequence of GenBank

accession number AB020878: nucleotides 35 894–36 958

and 56 217–56 362 encode human PAPP-A residues 59–413

and 414–461 (corresponding to murine residues 57–411 and

412–459, see Fig.1).Thus, the point of insertion of the

29-residue insert of murine PAPP-Ai corresponds to a junction

between two exons of the human gene, and the nucleotides

36.959–56.216 (of AB020878) either (a) correspond to a single intron of the human gene, or (b) contain an exon corresponding to the 29-residue murine insert.Within this stretch of  20 000 nucleotides, no human sequence was found that, when translated, showed significant sequence similarity to the insert of murine PAPP-Ai.We further looked for ORFs that were flanked by donor and acceptor splice sites [32], and found five candidate stretches of 15–76 amino acids.None of these, however, matched the required exon phase.In addition, none of them were represented in the human subset of the GenBank expressed sequence tag (EST) database, whereas > 100 human EST sequences encoded a sequence spanning the exon/exon junction at residues 413/414.Based on this, the presence of an insert at this position is unique to the mouse.It has previously been found [33], using a human cDNA probe, that the murine genome contains only one PAPP-A gene.Thus, PAPP-A and PAPP-Ai mRNA likely results from alternative splicing

of a transcript from the same gene

Expression and functional analysis of murine PAPP-A and PAPP-Ai

The proteolytic domain of murine PAPP-A, as recently defined in the sequence of human PAPP-A [21] (Fig.1), does not deviate from the overall degree of conservation (89.7% of the 312 residues are conserved) All residues of the zinc binding consensus are conserved (Fig.1), which strongly suggests that the murine protein is also an active metalloproteinase.To experimentally verify this, full-length constructs encoding PAPP-A and PAPP-Ai were made, cloned into a mammalian expression vector, and used for transient transfection of 293T cells

The presence of recombinant murine PAPP-A in culture supernatants of transfected cells was then confirmed by the detection of proteolytic activity against IGFBP-4 and IGFBP-5 (Fig.2) Medium from cells transfected with empty vector did not have the ability to cleave IGFBP-4 (Fig.2A, lane 1), but medium from cells transfected with murine PAPP-A cDNA caused specific cleavage in the presence of added IGF-II (Fig.2A, lane 2).In the absence

of IGF-II, proteolysis was dramatically less pronounced (Fig.2A, lane 3).This highlights the enhancing effect of IGF on proteolysis of IGFBP-4, which is widely recognized for human PAPP-A [6].PAPP-Ai also specifically cleaved IGFBP-4 in an IGF-dependent manner (Fig.2, lanes 4 and 5).Interestingly, however, the amount of proteolysis (in the presence of IGF) appeared to be much lower when compared to PAPP-A (Fig.2A, lanes 2 and 4)

In a similar experiment, we found that both PAPP-A and PAPP-Ai were able to specifically cleave IGFBP-5 inde-pendent of IGF (Fig.2B).In contrast to the proteolysis of IGFBP-4, the presence of added IGF slightly hampered the proteolysis of IGFBP-5, as recently demonstrated with human PAPP-A [6]

To verify that both PAPP-A and PAPP-Ai are expressed

as full-length proteins, we performed Western blotting using polyclonal antibodies against the human PAPP-A/proMBP complex, which were found to recognize murine PAPP-A and PAPP-Ai immobilized on a PVDF membrane.This experiment demonstrates that both species are in fact expressed as dimers of  400 kDa (Fig.3), as human PAPP-A

The basic insert of PAPP-Ai restricts proteolysis

of IGFBP-4, but not IGFBP-5

Rates of proteolysis cannot be compared using a fixed time

of incubation, as in the experiment described above, where

samples were incubated for 24 h (Fig.2) During this

incubation, only a fraction of IGFBP-4 had been degraded

by PAPP-Ai (Fig.2, lane 4) In contrast, the almost

complete proteolysis by PAPP-A (Fig.2, lane 2) may have

occurred in much less time

To more accurately describe this apparent difference in

activity of PAPP-A and PAPP-Ai towards IGFBP-4,

analyses were carried out using a recently developed

quantitative assay [6].Samples were removed from the

reactions at several different time points, and the degree of

cleavage was determined by SDS/PAGE followed by

measurement of band intensities with aPHOSPHORIMAGER

This revealed that in comparison with PAPP-A, PAPP-Ai is

much less proteolytically active against IGFBP-4.As

measured after 180 min of incubation, the activity of

PAPP-Ai is only about 10% of the activity of PAPP-A

(Fig.4A)

A similar time course experiment was performed with

IGFBP-5 as the substrate.Surprisingly, PAPP-A and

PAPP-Ai degraded IGFBP-5 with very similar rates

(Fig.4B).We therefore conclude that the basic insert of

29 residues carried by PAPP-Ai differentially affects its

substrate specificity; PAPP-A proteolysis of IGFBP-5 is not

affected by the presence of the insert, but the ability of PAPP-A to cleave IGFBP-4 is dramatically reduced

mRNA species encoding both PAPP-A variants are present in several tissues

To verify the existence of both PAPP-A and PAPP-Ai mRNA in murine tissues, RT-PCR analysis was carried out using PCR primers spanning the site of insertion in the nucleotide sequence.Most of the tissues analyzed contained both mRNA species; in general, PAPP-A mRNA appeared

to be the most abundant (Fig.5A).Of particular interest, although the method does not allow quantitative compar-isons between tissues, expression of PAPP-A and PAPP-Ai mRNA in the murine placenta appeared similar to levels in other tissues analyzed.The expression of PAPP-A mRNA

in the human placenta, in contrast, exceeds expression in other human tissue by > 250-fold [34]

To experimentally verify the absence of a human transcript encoding an insert between residues 413 and

414, RT-PCR with primers derived from the corresponding part of the human PAPP-A sequence was also carried out using cDNA derived from human placenta as a template (Fig.5B).No band of increased size was seen, providing

Fig 4 Degradation of IGFBP-4 and IGFBP-5 by murine PAPP-A and PAPP-Ai as a function of time Recombinant murine PAPP-A (s and PAPP-Ai (d) (both at 0.1 n M ) were incubated with radiolabeled IGFBP-4 (144 n M ) (A) or IGFBP-5 (130 n M ) (B) in the presence of added molar excess of IGF-II.Samples of the reaction mixtures were taken at various time points, and the degree of cleavage was determined

by densiometry using a PhosphorImager after separation by SDS/ PAGE.Values are average of three independent experiments ± SD.

Fig 3 Western blotting of murine PAPP-A and PAPP-Ai Culture

supernatants from cells transfected with murine PAPP-A cDNA

(lane 1), empty vector (lane 2), or PAPP-Ai cDNA (lane 3) were

separated by nonreducing SDS/PAGE and blotted onto a

poly(vinylidene flouride) membrane for immunodetection.Polyclonal

antibodies against human PAPP-A/proMBP were found to be effective

in this procedure, thus recognizing the denatured murine PAPP-A and

PAPP-Ai.Positions of molecular mass markers are indicated.

further evidence for the lack of a human counterpart of the

murine PAPP-Ai variant

Murine pregnancy serum does not contain proteolytic

activity against IGFBP-4

We finally looked for IGFBP-4 proteolytic activity in

murine pregnancy serum (Fig.6).Even under conditions of

prolonged incubation, neither nonpregnant (Fig.6, lanes

1–2) nor pregnant (Fig.6, lanes 3–4) murine serum showed

any ability to convert IGFBP-4 into fragments

character-istic of PAPP-A proteolysis.However, two very faint

different bands seen with nonpregnant serum, but not

pregnant serum, indicated the possible presence in

non-pregnant murine serum of a different proteinase with very

limited ability to cleave IGFBP-4.Human nonpregnancy

serum did not show any cleavage of IGFBP-4 (Fig.6,

lane 5), but human pregnancy serum showed the expected

cleavage caused by PAPP-A (Fig.6, lane 6–7)

To exclude the possibility that the lack of PAPP-A

activity in murine pregnancy serum was caused by an

unknown inhibitor, we compared proteolysis of IGFBP-4

by recombinant murine PAPP-A in the absence and in the presence of added murine pregnancy serum (not shown)

No difference in activity was seen, supporting the conclusion that murine pregnancy serum does not contain PAPP-A

D I S C U S S I O N

We have cloned a cDNA encoding murine PAPP-A of 1545 residues, and we have identified a cDNA encoding a variant, PAPP-Ai, in which 29 residues are inserted in the proteolytic domain.Through expression in mammalian cells, we show that both PAPP-A and PAPP-Ai are active proteinases of about 400 kDa.Further analyses demonstrate that (1) both PAPP-A and PAPP-Ai cleave IGFBP-4 in an IGF dependent manner, but that PAPP-Ai is a much slower IGFBP-4 proteinase than PAPP-A (2) in contrast, both PAPP-A and PAPP-Ai cleave IGFBP-5 independent of IGF at very similar rates (3) mRNA encoding PAPP-A and PAPP-Ai are both present in most murine tissues analyzed, and (4) murine pregnancy serum does not possess an elevated level of proteolytic activity against IGFBP-4, in striking contrast to human pregnancy serum

As PAPP-A is abundantly expressed in the human placenta [13,34], we unexpectedly found only two partial PAPP-A cDNA clones upon hybridization with human cDNA to a murine placental cDNA library.The remaining sequence (the N-ternimal 498 residues) was obtained by a PCR procedure using specifically primed placental cDNA

as the template.Of the 1545 residues of murine PAPP-A, 91.1% are also found in human PAPP-A The two sequences can be aligned without introducing gaps, except for missing residues at two single positions (corresponding

to human residues 6 and 27).Critical residues, such as the 82 cysteine residues, residues of the elongated zinc-binding consensus, and the Met-turn residues, are conserved between mouse and man (Fig.1).Searching the GenBank database revealed 8 murine EST sequences derived from PAPP-A mRNA (none of these originate from placenta), and two partial cDNA sequences (AF260433 and AF258461) encoding murine PAPP-A, but lacking  500 nucleotides corresponding to residues 1 though 178 Because of the limited number of available matching EST clones, RT-PCR was performed on a series of murine tissues (Fig.5A).A set of primers was selected that tested for the

Fig 6 Comparison of IGFBP-4 proteolyticac tivity in murine and human pregnancy serum Radiolabeled IGFBP-4 was incubated (72 h) with serum from nonpregnant female mice (lanes 1–2), serum from near term (18 days) pregnant mice (lanes 3–4), serum from a nonpregnant woman (lane 5), and two different samples of term human pregnancy serum (lanes 6–7).All reactions were in the presence of added IGF–II The positions of intact and cleaved IGFBP-4 are indicated.

Fig 5 RT-PCR analysis of murine PAPP-A and PAPP-Ai mRNA.

(A) A panel of cDNA preparations derived from murine tissues was

screened by PCR for the absence or presence of mRNA encoding

PAPP-A and PAPP-Ai, respectively.The presence of both mRNA

species in several of the tissues analyzed is evidenucleotides Individual

tissues tested and bands of 177 and 264 bp, corresponding to PAPP-A

and PAPP-Ai mRNA, respectively, are indicated.(B) A similar

experiment using equivalent primers derived from the human PAPP-A

cDNA sequence and template derived from human placenta.No band

corresponding to the murine 264 bp band was observed.For

com-parison, the PCR products obtained with murine A and

PAPP-Ai cDNA are also shown in the lanes labeled murine PAPP-A and

murine PAPP-Ai.

presence of both PAPP-A and PAPP-Ai mRNA at the same

time.Most of the tissues analyzed were found to contain

both species.Although the assay used is not quantitative, it

is fair to conclude that expression in the murine placenta

does not differ dramatically from other tissues analyzed

This is in accordance with the above findings, but in striking

contrast to semiquantitative analyses of PAPP-A mRNA

expression in human tissues, which revealed that expression

in the human placenta exceeds expression in other tissue

250- to 3000-fold [34].In the human placenta, PAPP-A

mRNA is abundantly synthesized in the

syncytiotropho-blast [13], the chorionic epithelium of fetal origin which is in

direct contact with the maternal blood.Based on this direct

contact, the placenta of man (and other primates) and the

placenta of mouse (and other rodents) are classified together

as hemochorial.In contrast, the placentas of horses, pigs,

ruminants, cats and dogs etc.are of different types with

more separating layers.Thus, most likely, the synthesis of

PAPP-A does not correlate with placental type.The

detected PAPP-A mRNA of the murine placenta may

originate from cells of fetal or maternal connective tissue

The relatively high levels of PAPP-A circulating in

human pregnancy serum most likely originate from the

placenta.As PAPP-A mRNA expression in the murine

placenta is not elevated compared to other tissues, we did

not expect to find elevated levels of IGFBP-4 proteolytic

activity in late mouse pregnancy serum, which was

confirmed (Fig.6) Previously, the presence of intact

IGFBP-4 (identified as a band of 24 kDa by Western

ligand blotting) in murine late pregnancy serum provided

indirect evidence that an IGFBP-4 proteinase was absent

from the circulation of mouse [35], although human

pregnancy serum contained detectable intact IGFBP-4 only

before gestational week 10 [36].The lack of

IGFBP-4-specific proteolysis in murine pregnancy serum, as found

here, is thus in line with the earlier findings, but it could not

be ruled out previously that the apparently constant level of

intact IGFBP-4 in murine pregnancy was caused by an

increase in synthesis along with increased proteolysis

The presence of PAPP-A mRNA in all murine tissues

analyzed parallels the ubiquitous occurrence of PAPP-A in

human tissues.Several recent papers have reported

proteo-lytic activity against IGFBP-4 or -5 in conditioned media

from cultures of mouse or rat cells [23–28].Hence, PAPP-A

and PAPP-Ai are obvious candidate proteinases in these

systems.However, a murine homologue of PAPP-A2 may

also be responsible in part for proteolysis in these different

systems.Human PAPP-A2 was recently identified and

demonstrated to cleave IGFBP-5 [22].By searching the

GenBank database for murine nucleotide sequences

enco-ding protein similar to human PAPP-A2, we determined the

existence of this protein in mouse.Interestingly, the found

PAPP-A2 sequence stretches (AK005504, BB462397, and

AI157031, for example), showed a lower degree of

conser-vation (64–83% in stretches of 78–120 residues) than the

91% observed between human and murine PAPP-A

Our cloning of cDNA encoding both PAPP-A and

PAPP-Ai allowed expression in mammalian cells and

functional analyses of the recombinant proteins.Of

partic-ular interest is the finding that PAPP-Ai does not readily

cleave IGFBP-4, and that, in contrast, A and

PAPP-Ai cleave IGFBP-5 with very similar rates.This immediately

suggests that proteolysis of IGFBP-4 might be regulated by

the control of PAPP-A/PAPP-Ai mRNA splicing.Both mRNA species are present in all murine tissues analyzed However, at the level of individual cells or cell types within the tissues, PAPP-A and PAPP-Ai mRNA may be differ-entially expressed

Sequence stretches similar to the 29-residue insert sequence was not found within the genomic sequence of human PAPP-A that potentially encodes a corresponding human insert.But a functional role of the insert of the murine PAPP-Ai is strongly suggested from the above experiments, even though the mechanism of its action cannot be predicted.Curiously, the site of insertion within the proteolytic domain of PAPP-A lies in close proximity to the cysteine residue which in the human PAPP-A/proMBP complex forms a disulfide bond to proMBP [15] (see Fig.1)

As mentioned above, proMBP functions as a proteinase inhibitor in the PAPP-A/proMBP complex [18], but whe-ther any mechanistic parallel on regulation of proteolytic activity can be drawn between the insert of PAPP-Ai and the linkage to proMBP is not known.A striking common feature of the 29-residue insert and MBP (of 117 residues) is their pronounced basic characters, which may be important for the basis of their actions

Even though the pregnancy protein PAPP-A is practi-cally absent from the murine placenta, the mouse may prove useful for the study of physiological roles of PAPP-A outside this tissue.Development of the mouse as a model for the study of PAPP-A must take into account the existence of PAPP-Ai.In contrast to the human system, specific proteolytic activity against IGFBP-4 will depend on whe-ther PAPP-A or PAPP-Ai functions in a given system Importantly, the availability of an expression system for recombinant murine PAPP-A will allow generation anti-bodies against murine PAPP-A, highly desired for efficient use of a murine model, as well as detailed mapping of monoclonal antibodies by homology substitution.Further, murine PAPP-A is now available for biochemical studies of the fifth metzincin family, the pappalysins

A C K N O W L E D G E M E N T S

This work was supported by grants from the Danish Medical Research Council, the Alfred Benzon Foundation, the Novo Nordic Foundation, and the National Institute of Health (HD31579-07).

R E F E R E N C E S

1.Daughaday, W H.& Rotwein, P.(1989) Insulin-like growth factors I and II.Peptide, messenger ribonucleic acid and gene structures, serum, and tissue concentrations Endocrinol Rev 10, 68–91.

2 Hwa, V., Oh, Y & Rosenfeld, R.G (1999) The insulin-like growth factor-binding protein (IGFBP) superfamily Endocrinol Rev 20, 761–787.

3 Jones, J.I & Clemmons, D.R (1995) Insulin-like growth factors and their binding proteins: biological actions Endocrinol Rev 16, 3–34.

4.Fowlkes, J L.(1997) Insulinlike growth factor-binding protein proteolysis.An emerging paradigm in insulinlike growth factor physiology Trends Endocrinol Metab 8, 299–306.

5 Lawrence, J B , Oxvig, C , Overgaard, M T , Sottrup-Jensen, L , Gleich, G.J., Hays, L.G., Yates III, J.R & Conover, C.A (1999) The insulin-like growth factor (IGF)-dependent IGF binding protein-4 protease secreted by human fibroblasts is

pregnancy-associated plasma protein-A Proc Natl Acad Sci USA 96, 3149–

3153.

6 Laursen, L.S., Overgaard, M.T., Soe, R., Boldt, H.B.,

Sottrup-Jensen, L., Giudice, L.C., Conover, C.A & Oxvig, C (2001)

Pregnancy-associated plasma protein-A (PAPP-A) cleaves

insulin-like growth factor binding protein (IGFBP)-5 independent of

IGF: implications for the mechanism of IGFBP-4 proteolysis by

PAPP-A FEBS Lett 504, 36–40.

7 Kristensen, T , Oxvig, C , Sand, O , Moller, N P &

Sottrup-Jensen, L.(1994) Amino acid sequence of human

pregnancy-associated plasma protein-A derived from cloned cDNA.

Biochemistry 33, 1592–1598.

8 Qin, X , Byun, D , Strong, D D , Baylink, D J & Mohan, S.

(1999) Studies on the role of human insulin-like growth factor-II

(IGF-II) - dependent IGF binding protein (hIGFBP) -4 protease

in human osteoblasts using protease-resistant IGFBP-4 analogs.

J Bone Miner Res 14, 2079–2088.

9 Bayes-Genis, A., Schwartz, R.S., Lewis, D.A., Overgaard, M.T.,

Christiansen, M , Oxvig, C , Ashai, K , Holmes Jr, D R &

Con-over, C.A (2001) Insulin-like growth factor binding protein-4

protease produced by smooth muscle cells increases in the

cor-onary artery after angioplasty Arterioscler Thromb Vasc Biol.

21, 335–341.

10 Bayes-Genis, A., Conover, C.A., Overgaard, M.T., Bailey, K.R.,

Christiansen, M., Holmes Jr, D.R Virmani, R., Oxvig, C &

Schwartz, R.S (2001) Pregnancy-associated plasma protein A as a

marker of acute coronary syndromes N Engl J Med 345, 1022–

1029.

11 Conover, C.A., Oxvig, C., Overgaard, M.T., Christiansen, M &

Giudice, L.C (1999) Evidence that the insulin-like growth factor

binding protein-4 protease in human ovarian follicular fluid is

pregnancy associated plasma protein-A J Clin Endocrinol.

Metab 84, 4742–4745.

12 Conover, C A , Faessen, G F , Ilg, K E , Chandrasekher, Y A ,

Christiansen, M., Overgaard, M.T., Oxvig, C & Giudice, L.C.

(2001) Pregnancy-associated plasma protein-a is the insulin-like

growth factor binding protein-4 protease secreted by human

ovarian granulosa cells and is a marker of dominant follicle

selection and the corpus luteum Endocrinology 142, 2155–2158.

13 Bonno, M., Oxvig, C., Kephart, G.M., Wagner, J.M., Kristensen,

T., Sottrup-Jensen, L & Gleich, G.J (1994) Localization of

pregnancy-associated plasma protein-A and colocalization of

pregnancy-associated plasma protein-A messenger ribonucleic

acid and eosinophil granule major basic protein messenger

ribo-nucleic acid in placenta Laboratory Invest 71, 560–566.

14 Oxvig, C , Haaning, J , Kristensen, L , Wagner, J M , Rubin, I ,

Stigbrand, T., Gleich, G.J & Sottrup-Jensen, L (1995)

Identifi-cation of angiotensinogen and complement C3dg as novel proteins

binding the proform of eosinophil major basic protein in human

pregnancy serum and plasma J Biol Chem 270, 13645–13651.

15 Oxvig, C., Sand, O., Kristensen, T., Gleich, G.J & Sottrup-Jensen,

L.(1993) Circulating human pregnancy-associated plasma

pro-tein-A is disulfide-bridged to the proform of eosinophil major

basic protein J Biol Chem 268, 12243–12246.

16 Oxvig, C., Sand, O., Kristensen, T., Kristensen, L &

Sottrup-Jensen, L.(1994) Isolation and characterization of circulating

complex between human pregnancy-associated plasma protein-A

and proform of eosinophil major basic protein Biochim Biophys.

Acta 1201, 415–423.

17 Popken-Harris, P , Checkel, J , Loegering, D , Madden, B ,

Springett, M., Kephart, G.& Gleich, G.J.(1998) Regulation and

processing of a precursor form of eosinophil granule major basic

protein (ProMBP) in differentiating eosinophils Blood 92,

623–631.

18 Overgaard, M.T., Haaning, J., Boldt, H.B., Olsen, I.M., Laursen,

L.S., Christiansen, M., Gleich, G.J., Sottrup-Jensen, L., Conover,

C A.& Oxvig, C.(2000) Expression of recombinant human

pregnancy-associated plasma protein-A and identification of the proform of eosinophil major basic protein as its physiological inhibitor J Biol Chem 275, 31128–31133.

19 Stocker, W.& Bode, W.(1995) Structural features of a super-family of zinc-endopeptidases: the metzincins Curr Opin Struct Biol 5, 383–390.

20 Stocker, W , Grams, F , Baumann, U , Reinemer, P , Gomis-Ruth, F.X., McKay, D.B & Bode, W (1995) The metzincins – topological and sequential relations between the astacins, adamalysins, serralysins, and matrixins (collagenases) define a superfamily of zinc-peptidases Protein Sci 4, 823–840.

21 Boldt, H B , Overgaard, M T , Laursen, L S , Weyer, K , Sottrup-Jensen, L.& Oxvig, C.(2001) Mutational analysis of the proteolytic domain of pregnancy-associated plasma protein-A (PAPP-A): classification as a metzincin Biochem J 358, 359–367.

22 Overgaard, M.T., Boldt, H.B., Laursen, L.S., Sottrup-Jensen, L., Conover, C.A.& Oxvig, C.(2001) Pregnancy-associated plasma protein-A2 (PAPP-A2), a novel insulin-like growth factor-binding protein-5 proteinase J Biol Chem 276, 21849–21853.

23 Chernausek, S.D., Smith, C.E., Duffin, K.L., Busby, W.H., Wright, G.& Clemmons, D R.(1995) Proteolytic cleavage of insulin-like growth factor binding protein 4 (IGFBP-4).Locali-zation of cleavage site to non-homologous region of native IGFBP-4 J Biol Chem 270, 11377–11382.

24 Thrailkill, K.M., Quarles, L.D., Nagase, H., Suzuki, K., Serra, D.M & Fowlkes, J.L (1995) Characterization of insulin-like growth factor-binding protein 5-degrading proteases produced throughout murine osteoblast differentiation Endocrinology 136, 3527–3533.

25 Fowlkes, J.L., Serra, D.M., Rosenberg, C.K & Thrailkill, K.M (1995) Insulin-like growth factor (IGF) -binding protein-3 (IGFBP-3) functions as an IGF-reversible inhibitor of IGFBP-4 proteolysis J Biol Chem 270, 27481–27488.

26 Fowlkes, J.L., Thrailkill, K.M., George-Nascimento, C., Rosen-berg, C.K & Serra, D.M (1997) Heparin-binding, highly basic regions within the thyroglobulin type-1 repeat of insulin-like growth factor (IGF)-binding proteins (IGFBPs)-3–5, and -6 inhibit IGFBP-4 degradation Endocrinology 138, 2280–2285.

27 Resnick, C.E., Fielder, P.J., Rosenfeld, R.G & Adashi, E.Y (1998) Characterization and hormonal regulation of a rat ovarian insulin-like growth factor binding protein-5 endopeptidase: an FSH-inducible granulosa cell-derived metalloprotease Endocrinology 139, 1249–1257.

28 Smith, E P , Kamyar, A , Niu, W , Wang, J , Cercek, B , Chernausek, S.D & Fagin, J.A (2001) IGF-binding protein-4 expression and IGF-binding protein-4 protease activity are regu-lated coordinately in smooth muscle during postnatal develop-ment and after vascular injury Endocrinology 142, 4420–4427.

29 Ho, S N , Hunt, H D , Horton, R M , Pullen, J K & Pease, L R (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction Gene 77, 51–59.

30 DuBridge, R.B., Tang, P., Hsia, H.C., Leong, P.M., Miller, J.H.

& Calos, M.P (1987) Analysis of mutation in human cells by using an Epstein-Barr virus shuttle system Mol Cell Biol 7, 379–387.

31 Pear, W.S., Nolan, G.P., Scott, M.L & Baltimore, D (1993) Production of high-titer helper-free retroviruses by transient transfection Proc Natl Acad Sci USA 90, 8392–8396.

32 Solovyev, V.V., Salamov, A.A & Lawrence, C.B (1994) Pre-dicting internal exons by oligonucleotide composition and dis-criminant analysis of spliceable open reading frames Nucleic Acids Res 22, 5156–5163.

33 Pilz, A , Woodward, K , Povey, S & Abbott, C (1995) Com-parative mapping of 50 human chromosome 9 loci in the laboratory mouse Genomics 25, 139–149.

34 Overgaard, M.T., Oxvig, C., Christiansen, M., Lawrence, J.B., Conover, C.A., Gleich, G.J., Sottrup-Jensen, L & Haaning, J.

(1999) Messenger ribonucleic acid levels of pregnancy-associated

plasma protein-A and the proform of eosinophil major basic

protein: expression in human reproductive and nonreproductive

tissues Biol Reprod 61, 1083–1089.

35 Fielder, P.J., Thordarson, G., Talamantes, F & Rosenfeld, R.G.

(1990) Characterization of insulin-like growth factor binding

proteins (IGFBPs) during gestation in mice: effects of

hypophy-sectomy and an IGFBP-specific serum protease activity Endocrinology 127, 2270–2280.

36 Giudice, L.C., Farrell, E.M., Pham, H., Lamson, G & Rosenfeld, R.G (1990) Insulin-like growth factor binding proteins in maternal serum throughout gestation and in the puerperium: effects of a pregnancy- associated serum protease activity J Clin Endocrinol Metab 71, 806–816.